the shipwrecks of heracleion-thonis: an overview · the shipwrecks of heracleion-thonis: an...

HAL Id: halshs-00845538https://halshs.archives-ouvertes.fr/halshs-00845538

Submitted on 17 Jul 2013

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

The Shipwrecks of Heracleion-Thonis: An OverviewDavid Fabre, Alexander Belov

To cite this version:David Fabre, Alexander Belov. The Shipwrecks of Heracleion-Thonis: An Overview. Achievementsand problems of modern Egyptology. Proceedings of the international conference. September 29-October 4, 2009, Moscow, Sep 2009, Moscow, Russia. pp.107-118. �halshs-00845538�

https://halshs.archives-ouvertes.fr/halshs-00845538

https://hal.archives-ouvertes.fr

Russian Academy of Sciences

Center for Egyptological Studies

ACHIEVEMENTS AND PROBLEMS OF

MODERN EGYPTOLOGY

Proceedings of the International Conference

Held in Moscow on September 29–October 2, 2009

Edited by Galina A. Belova

Moscow 2011

The Shipwrecks of Heracleion-Thonis: An Overview 107

Topographical surveys and soundings in the Bay of Aboukir led to the discovery of

the Heracleion–honis site and allowed for an initial understanding of the religious, economical

and urban organisation of the site. he results of the irst ten years of excavations (1996–2006)

are published in a separate monograph5 and here we will consider only those peculiarities of

honis-Heracleion’s topography that shed some light on the position of the shipwrecks discov-

ered at the site.

2. General information on the shipwrecks

2.1. Topography of the Site

he site of Heracleion-honis covers an area of 1000 by 1200 m. he most impressive

constructions on the site, including the Temple, are situated on the peninsula that was separated

from the Nile by a Southern Basin and a narrow passage between the sand dunes. he Central

Basin, North Basin and Western Lake surround the central part of the city on three sides. It

is worth remembering that the toponym of honis is believed to originate from the name of

the coastal artiicial lake (Hn.t).6 A large canal crosses the peninsula from the Central Basin to

the Western Lake in the vicinity of the temple.7 he complex of the city is well protected from

the dominating north-western wind. hese are the main features of the city while its actual to-

pography is complicated by secondary canals and passages.

Numerous remains of quays have been discovered while the bottom of the port areas and

canals are very clearly indicated by the presence of ancient anchors of diferent types. More than

700 of them have been found so far bearing witness to the great activity of the port. Among

these, a speciic type of a stone anchor with wooden lukes is preponderant but there are also

considerable amounts of stone and lead stocks from wooden anchors. Usually it is not possible

to date these anchors without good context as they remained in use over many centuries. How-

ever, wooden lukes have been preserved for some of the anchors discovered so far thus allowing

carbon-dating them between the middle of the 6th century bc and the end of the 4th century

bc. Apart from the anchors, navigation areas are distinguished by sediment deposits which are

very rich in pottery sherds and small fragments of lead. he area of the Grand Canal, apart from

anchors, is characterised by numerous ritual and votive objects.

he information provided by the distribution of the anchors is of major interest and de-

serves attention but the main archaeological potential of the port areas of Heracleion lies in their

great number of ancient shipwrecks.8 More than 60 ships were identiied, an accumulation never

encountered before at any other archaeological site.9 he preliminary studies of the shipwrecks

5 Goddio, Topography and Excavation. See also J.-D. Stanley, Geoarchaeology: Underwater Archaeology in the Canopic Region in Egypt (Oxford, 2007).

6 J. Yoyotte, ‘Notes de toponymie égyptienne» in Festschrift zum 80. Geburgstag von Professor Dr. Hermann Junker’ in MDAIK 16 (1958), II, 414–430.

7 Goddio, Topography and Excavation, 74–75.8 D. Fabre, ‘he shipwrecks of Heracleion-

honis. Preliminary Study and Research Perspectives’ in Wilson A., Robinson D. (eds.) East meets West

along the Maritime Silk Route. Waseda University. Tokyo 2–3 July 2009. Oxford, forthcoming.

9 he number of Roman ships found in Pisa, Italy, in 1999 amounts to 16. See S. Bruni (ed.), Le navi antiche de Pisa: ad un anno dall’inizio delle recherche. he Ancient Ships of Pisa: after a year of work (Florence, 2000). he number of Yenikapi wrecks in Turkey run to 34 ships dating from 5th to 11th century ad, See U. Kocabaş, ‘Byzantine shipwrecks project’ in Byz-antium and the Sea, conference held in Vienne 19-21 February 2010, forthcoming.

David Fabre, Alexander Belov108

which consisted in determining their shapes, sizes, the wood species they were made of and their

date showed that, in most cases, the degree of the wood’s conservation is very good, sometimes

exceptional. he good state of preservation of the material is due to the presence of a very dense

clay layer that covers considerable parts of many of the hulls. On the downside, this clay layer

slows down signiicantly the excavation process.

2.2. Dating of the Shipwrecks10

As the quantity of material is overwhelming, it was not possible, in the irst stage, to collect

more than a few samples from each shipwreck, and thus dates and especially paleobotanic stud-

ies, are still subject to change. However, the irst results indicate that the majority of shipwrecks

date to between the 6th and the 2nd centuries bc. Forty ive shipwrecks that have calibrated radio-

carbon dates are presented on Figure 1 and some preliminary conclusions can be drawn at this

point. We can distinguish ive chronological groups for the ships. he irst includes seven ship-

wrecks that deinitively belong to the Ptolemaic period. Eleven ships that most probably belong

to the same period form the second group. Dating of eighteen ships falls into the Late Period

(664–332 bc). Another eight ships can belong either to the Late or to the hird Intermediate

Period (1069–664 bc). he last group is represented by a single shipwreck which dates from

the 14th to the 11th century bc and is thus of special interest.

To summarize 40% of the ships are most probably Ptolemaic in date. Although the Greco-

Roman type of construction is well known from numerous excavations in the Mediterranean,

ships of the Ptolemaic period have never been excavated in Egypt so far. he large number of

ships of this period at Heracleion can be used to illustrate some peculiar features of construc-

tion of purely local origin and testiies to the use of a selection of indigenous wood species.

On the other hand, good conservation of the material can help testing unproved hypotheses

concerning the Greco-Roman boatbuilding tradition. Another 40% of the shipwrecks belong

to the Late Period. Very few shipwrecks of this age were excavated throughout the Mediter-

ranean and almost each new one brings precious information. As far as Egypt is concerned,

the lack of such type of archaeological data for this period is especially striking and studies of

naval architecture are mainly based on scarce iconographic and epigraphic material. However,

the Late Period is characterised by numerous technological improvements in naval architec-

ture. he study of the transition from the indigenous Egyptian boatbuilding tradition to the

Greco-Roman one opens up avenues of investigation. It is quite possible that some of the eight

ships that are dated between the end of the 9th and the end of the 5th century bc actually belong

to the hird Intermediate Period and as such, may bear witness to constructional features that

appeared only during the time of the Persian domination. Finally, one ship has a calibrated

dating falling between the 14th and the end of the 11th century bc. Such discovery is absolutely

exceptional as only few examples of underwater excavations of ships of this age can be cited.

10 We draw attention to the fact that the range of dates suggested by radio carbon is quite large, even after adjustments have been made for environmental factors (plus or minus 50 years, or even more). Further-more, the date derived from C14 methods only relates

to the period during which the wood was living mat-ter, that is, when it was still sapwood. he data, with C14 analyses references, will be published in a mono-graph devoted to the shipwrecks of Heracleion-honis (OCMA Monograph, Oxford).


In addition, neither the Phoenician shipwreck of Uluburun (1300 bc)11 nor the Cape Gelidonia

shipwreck (1200 bc)12 show many constructional details.

2.3. Distribution of the Shipwrecks

he shipwrecks form zones of accumulation that are located mainly in the Central Basin

and in the Grand Canal. Two major accumulations in the western and eastern part of the Cen-

tral Basin include 11 ships each. he main conclusion that one can draw looking at Figure 2

is that the majority of the shipwrecks in each accumulation form synchronous groups. hus

probably, the accumulations were not formed by natural factors acting over long periods of time,

such as tides or currents, but by some sudden historical events. However, the latter could be

either of natural origin (for example earthquake with resulting swell or hurricane) or of human

making (sinking of a leet during a war or for some speciic purpose in peaceful times). Still, the

context of the shipwrecks makes us believe that natural hazard is much less probable than hu-

man activity, as will be shown below.

3. Preliminary Studies of Naval Architecture

3.1. Wood Species Used in the Construction of the Ships

Two groups of wood species are attested as building material for the construction of

the ships: those that are native to Egypt and those that could have been imported.

he natural dearth of large trees in Egypt forced the local shipwrights since prehistor-

ic times to make use of the native poor quality wood in spite of its numerous disadvantages.

At the same time, it is this factor that stimulated the inventiveness of the Egyptians and led to

the appearance of their unique boatbuilding tradition.

Acacia (Acacia sp.). Acacia wood dominates as a construction material for the ships discov-

ered in Heracleion as almost 80% of them have at least some strakes of planking made of acacia

(Acacia sp., Acacia totilis/radiana). he following table and Figure 3 illustrate the number and

percentage of ships as per the wood species used for their construction.

Wood Number of Ships

Acacia 47

Sycomore 1

Pine 2

Oak 4

Undetermined 6

Total 60

Keeping in mind that acacia is a very peculiar wood and that nowhere in the Mediterra-

nean could it have been considered as appropriate material for boatbuilding, one may presume

that these ships probably had all their planking or even their entire construction made of acacia.

11 C. Pulak, ‘he Uluburun Shipwreck, an Update’ in 6th International Symposium on Ship Construction in Antiquity Proceedings (Tropis VI, Athens, 2001), 439.

12 G.F. Bass, Cape Gelidonya: A Bronze Age Ship-wreck (Transactions of the American Philos. Society New Series 57, Philadelphia, 1967).


Acacia wood is very diicult to work. Herodotus, however, left us the description of the con-

struction of the authentic Egyptian river craft baris, built of acacia.13 heophrastus and Pliny the

Elder mention acacia among the trees used for boatbuilding in Egypt and describe a region near

hebes where this wood was worked.14 During the Ptolemaic and Roman Periods, there were

still acacia groves around Lake Sirbonis in the eastern Delta.15 It is probable that acacia forests

also existed in the Canopic region.

Sycomore (Ficus sycomorus). he other local wood found in the boats construction is syco-

more. It is a poor quality wood but, nevertheless, it was used in Egypt since the Neolithic era.

Documents of the 18th Dynasty and of the Ptolemaic period prove that sycomore was used for

boatbuilding.16 he irst archaeological evidence for it comes from the boat found at Matariya

(near Heliopolis) in 1987. his river boat dated to the 5th century bc was 11 m long.17 At least

one strake of shipwreck 11 from the Grand Canal of Heracleion was cut of Ficus sycomorus. his

shipwreck has not yet been excavated and dates from the 2nd century bc. Moreover, the excava-

tions of the Roman ship of the Island of Antirhodos (1st century bc–1st century ad) in the Portus

Magnus of Alexandria, showed that one of its aft loor-timber was cut of sycomore. Both ves-

sels are not boats but ships of considerable size, the latter exceeding 30 m in length.18 Likewise,

Romans were using the wood of sycomore’s relative — Ficus carica — for boat construction as

evidenced by the recent excavation of the large Roman oared vessel from Pisa (Ship C).19

Pine and Oak. It is not possible at the present stage of excavation to ascertain whether

the oak and pine wood identiied on the Heracleion shipwrecks was imported or whether it was

of local origin. Wood importation intensiied during the 18th Dynasty following the deep pen-

etration of the Egyptians into Asia Minor. Oak, together with other species like maple, ash, elm,

olive and others, was imported to Egypt. Nethertheless, the texts by heophrastus and Pliny

the Elder show that oak was naturally present in the hebaid. he habitat of the Aleppo pine

(Pinus halepensis) also included some regions of North Africa.

Four shipwrecks of Heracleion have planking of oak. It should be noted that classical Medi-

terranean boatbuilding tradition very reluctantly resorted to oak when creating the shell of the ship

usually preferring resinous species and especially pine.20 Most often, oak was used either for the keel

and false keel or for transversal details, like frames and beams. It is notoriously diicult to change

the natural form of oak wood and thus it was used for the planking either intentionally in separate

strakes, like whales, or because of the absence of any other suitable material.21 If we suppose that

the wood for our ships was imported, would it not have been preferable to order softer wood, say

pine, which is easy to work and which is sea friendly? Ancient Egyptian shipwrights had a long

experience in working with resinous wood as we can see from the boat of Cheops.

13 Herodotus, Histories II, 96.14 Theophrastus, Enquiry into Plants II, 4, 2, 8;

Pliny, Natural History XIII, 63, 19.15 P. Chuvin and J. Yoyotte, ‘Documents rela-

tifs au culte pélusien de Zeus Casios’ in RevArch 1 (1986), 50.

16 R. Gale, P. Gasson and N. Hepper, ‘Wood’ in P. Nicholson and I. Shaw (eds.), Ancient Egyptian Materials and Technology (Cambridge, 2003), 340.

17 S. Vinson, Egyptian Boats and Ships (London, 1994), 47–48.

18 he results of the excavations of the Antirho-dos Island ship are being prepared for publication by IEASM (D. Fabre, A. Belov, P. Sandrin).

19 Bruni, Le navi antiche di Pisa, 83.20 M. Rival, La charpenterie navale romaine (Paris,

1991), 87.21 he shipwreck of Laurons (2nd cent. ad) had

a whale of oak; the wreck of Ladispoli (1st cent. bc) — some strakes. he planking of two Roman ships of Pisa was cut of oak. See Rival, La charpenterie nav-ale, 89; Bruni, Le navi antiche di Pisa, 81.


he construction of shipwreck 17 which was studied during the irst season of excava-

tions in spring 2009 shows features of indigenous construction. he keel of this ship consists

of 10 pieces of an average length of 2,4 m. Oak was identiied in the planking of the ship while

the texture and colour of its wood shows that the keel is probably of oak too. We have to con-

tinue our studies of this ship in order to understand whether the shipwrights initially possessed

only short saw-timber — and thus we may presume a local, African, origin for the wood — or

whether they intentionally used very short keel sections.

Pine was identiied in the construction of two ships of Heracleion. One of them is medi-

eval in date and can be excluded from our study. he wood of the other ship could have been

imported or the ship itself could have been of foreign origin. More astonishing is the absence

of cedar and cypress among the resinous wood species used for the construction of the ships.22

Phoenician wood had an excellent reputation for boatbuilding since the Predynastic period and

it kept it into the Achaemenid period. Cypresses from the island of the same name were used for

naval construction and the deliveries of wood from Cyprus to Egypt are mentioned in the El-

Amarna letters (14th century bc). Cedar from Lebanon was identiied also in the fragments of

the planking of the seagoing ship from the Middle Kingdom harbor at Wadi Gawasis on the

Red Sea.23

However, we should be cautious in reconstructing the direction of trade lows as a distinc-

tion need be drawn between construction wood (raw material) and pieces that have already been

worked. his diference is already present in the customs accounts of the satrapy of Egypt24, in

the Stele of Naukratis25 and in the Stele of Heracleion-honis.26

3.2. Elements of Construction

Keel. he keel of shipwreck 17 dating from the hird Intermediate or Late Period is made

up of 10 pieces and only one of them exceeds 3 m in length. he fragments of the keel are joined

by a Z-scarf with the key. he width of the keel is almost constant and changes within limits of

39–43 cm. During the irst season of excavation the keel was not studied in its section and we

can only guess at its internal structure and form. Yet this wide keel consisting of short timbers

makes one think that the shipwrights did not confer signiicant importance to the structural role

of the keel and, quite probably, it was not a proper keel but just a keel plank. In fact, it appears

to be the same type of plank but wider and thicker than the others which, are positioned in the

center and used as the starting point for the entire construction. At least at both extremities

of the ship the thickness of the timbers was very modest. Keel planks are known from many

shipwrecks dating from the 6th–5th centuries bc27 and the same type of keel was found in the

22 Gale, Gasson and Hepper in Nicholson and Shaw, Ancient Egyptian Materials and Technology, 334–352, 349–350.

23 K.A. Bard and R. Fattovich (eds.), Harbor of the Pharaohs to the Land of Punt, Archaeological In-vestigations at Mersa/Wadi Gawasis Egypt, 2001–2005 (Naples, 2007).

24 B. Porten and A. Yardeni, Textbook of Aramaic Documents from Ancient Egypt, III, Literature, Accounts, Lists (Jerusalem, 1993); Briant and Descat in Grimal and Menu, Le commerce, 59–104.

25 M. Lichtheim, Ancient Egyptian Literature, III, he Late Period (Berkeley–Los Angeles–London, 1980), 88.

26 D. Fabre, ‘Heracleion-honis: Customs Sta-tion and Emporion’ in Goddio and Fabre, Egypt’s Sunken Treasures, 219–234.

27 S. McGrail, ‘Sea Transport. Part I: Ships and Navigation’ in J.P. Oleson (ed.), he Oxford Hand-book of Engineering and Technology in the Classical World (Oxford, 2008).


construction of the Bronze Age ship of Uluburun. Probably this keel is devoid of rabbet and the

garboard joins the keel edge-to-edge. No breaks were found elsewhere along the garboard. How-

ever, were the initial angle of the planking high (boarding rises quickly), the fractures and spaces

would be visible between the planking and the keel. Together with the other facts considered

below this provides us with some evidence for suggesting that the hull of the ship was rather full.

Planking. he carvel planking of all the ships that were found on the site of Heracleion

was assembled by mortise and tenon joinery. his kind of assemblage is most typical of Greco-

Roman shipbuilding. Earlier shipwrecks of the archaic period in Greek history (650–480 bc)

from the northern coast of the Mediterranean show many examples of completely28 or partially

sewn boats.29 he Egyptians knew the technique of the ‘sewn boats’ since the third millennium

B.C. and the best example of it is the funeral barque of Cheops. For the time being, we are not

able to conirm the complete absence of sewing in the construction of the Heracleion ships.

However, no sewing was evidenced on shipwreck 17 which is, for the moment, the only ship

which has been partially excavated.

Still, one can immediately see the diference between the Heracleion’s wrecks planking

assemblage and those of the classical boatbuilding tradition. In our ships belonging to the Late

Period the distance between mortises varies between 20 and 35 cm. Likewise, average distance

on the archaic shipwrecks of Place Jules-Verne 7 and César I in Marseilles (525–510 bc) was

also 20 cm,30 on the ship of the Bronze Age of Uluburun this distance was around 25 cm.31 It

is generally recognised that, at that time, mortise and tenon joinery did not achieve the op-

timal density that was used by shipwrights in later periods. For comparison, average mortise

to mortise distance of the hulls from the 4th century bc to the 3rd century ad is only 12,5 cm,

almost twice as much less than what can be observed on the Heracleion ships.32 Still, the idea

of this type of joinery was to provide maximum lateral strength to the hull where tenons served

as internal frames. In our case, very speciic tenons have been used (Fig. 4). hey are not small

and rectangular pieces of wood that join only adjacent strakes of planking, but long and thick

timbers that pierce several planks.

he joint between the planks of the same strake is also peculiar. Instead of the simple di-

agonal scarf used in Greco-Roman shipbuilding, we ind here a half-lap splice. he extremity of

each plank at the joint is supported by a tenon. In fact, planking usually has an extraordinary

thickness; it is almost rectangular in section. he majority of the planks have dimensions within

the following range: 12–17 cm wide and 10–15 cm thick. his is a consequence of the primary

role of the planking in the structure of the ship and of using massive tenons inside relatively

short planks. Some wider planks correspond to the whales intended to support the hull at

the lines of maximum stress, especially near the waterline.

Such a construction of the hull corresponds to the technique described by Herodotus

who visited Egypt in the 5th century bc According to him, Egyptian cargo boats were made of

28 L. Basch, ‘Le navire cousu de Bon-Porté’ in CAS V (1976), 37–42.

29 P. Pomey, ‘Les épaves grecques archaïques du VIe siècle av. J.-C. de Marseille: épaves Jules-Verne 7 et 9 et César I’ in 6th International Symposium on Ship Construction in Antiquity, 425–437.

30 Pomey in 6th International Symposium on Ship

Construction in Antiquity, 428.31 Pulak in 6th International Symposium on Ship

Construction in Antiquity, 439.32 R. Steffy, ‘A Mediterranean ship construction

database: dating and classifying shipwrecks by their hull remains’ in 6th International Symposium on Ship Construction in Antiquity, 547–562.


acacia wood33 that had been ‘sawed into planks two cubits long, which they used in the manner

of bricks’. A tomb painting of the second millennium bc from Beni Hassan shows the process

of assembling the hull of a boat with remarkably short planks.34 We have already found several

shipwrecks with planks not exceeding 100–105 cm in length (or two cubits; one cubit being

equivalent to 52 cm). However, some other ships are built with longer planks (up to 6 m in

the case of shipwreck 17 made of oak).

Remains of unidentiied organic material were found in the joints of the planking. Egyp-

tians sometimes used papyrus leaves to make their ships watertight. Liquid pitch or bitumen was

poured additionally to protect the seams from water. In any case, this organic matter was prob-

ably some kind of oakum that was not forced into the seams as per later tradition but simply

applied on the external surfaces of the seam. Forcing oakum into the seams would have afected

the structural integrity of the hull. he material is being analyzed.

Framing. According to our current knowledge, the ships of Heracleion were built in

a ‘shell-irst’ technique. his means that planking was added to the keel irst and framing was

inserted at a later stage of the construction. Some frames are very coarse and carelessly worked.

At the same time, it seems that, in a majority of cases, they are positioned symmetrically from

both sides of the keel. Treenails were used to attach the frames to the planking. Still, we do not

know where the framing was attached to the keel.

he excavation of shipwreck 17 shows that at least on some ships very interesting frames

have been used which are not known elsewhere. Four frames of impressive dimensions have been

discovered in a symmetric position across the keel. he width of these frames is in the range of

30–35 cm (sided) and their thickness is about 10 cm (molded). he surface of the frames is ac-

curately worked. heir dimensions are quite unusual, as well as the spacing between them which

reaches 4 m. However, additional frames of smaller size were inserted between the large ones.

Each large frame has one central mortise on top and several lateral mortises in the center and at

each of their extremities. Most probably, the frames were attached to the keel and had a continu-

ation outboard that was not preserved or was removed intentionally in antiquity.

Upper structures. he remains of the beams on several of the Heracleion’s shipwrecks attest

to the existence of a deck. he same inding is conirmed by the the presence of a hatch found

on shipwreck 17 and by numerous top mortises in the frames and the keel that most probably

housed the stanchions that supported the deck.

Form of the hull. Preliminary studies show that the majority of the ships had rather broad

hulls and relatively lat bottoms. hese characteristics are perfectly suited to the natural condi-

tions of the shallow coastal lake that existed at Heracleion.35 he average length of the ships

varies from 16 to 26 m. he length to breadth ratio is not high and is in the range of 2,2 to 2,8

which is characteristic of a merchant ship.

We already noted the very short timbers of the planking of the majority of the ships

and this fact is of outmost signiicance. If we are dealing with seagoing ships, as suggested by

the presence of stone anchors on board and in vicinity of the ships, then their hulls would re-

quire additional longitudinal support. Traditionally in Egyptian boats this support was provided

33 Herodotus, Histories II, 96. 34 L. Casson, Ships and Seamanship in the Ancient World (Princeton, 1971), ig. 11, 13.


by a hogging truss that passed from the stern to the bow above the deck along the central axis of

the ship. hese trusses are known from a number of representations. Girdling was another im-

portant feature that reinforced the structure of the ship. During our excavations, many remains

of vegetal trusses were found. However, judging by their modest size (5–7 cm), none of them

could have served as a hogging truss.

In several cases, the stem of the ship is preserved within the construction and will be

excavated and studied in due time. As far as the construction of the stern is concerned, a very

important detail has emerged during the excavations of shipwreck 17, namely a massive timber

at the stern of the ship which measured over 3 min length and possessed two through openings

with respective diameters of 33 and 39 cm. Most probably these openings correspond to the

wells of the steering oars known from a number of depictions.

It was probably the full beams, modest keels and comparatively lat bottoms of these ships

that required the use of the ballast that was found on some of the ships.36 So far, three difer-

ent types of ballast were identiied on the shipwrecks. he irst looks like a bulk load of pebble

that was found along the keel of some of the ships, though it is still debatable whether it should

be interpreted as ballast or not. he second type includes limestone rubble that was found in

a compact mass in a bilge area in suicient quantity to serve as ballast. However, quite frequently

it was cargo itself that ancients used as ballast and it can be the case here too. Finally, blocks of

limestone were discovered as well (Fig. 5).

It is diicult for the time being to evaluate the tonnage of the ships discovered. However,

taking into consideration their lat bottoms it could not have been high. An estimate can be

provided by the Papyrus of Elephantine which includes a customs register for Egypt from the

Achaemenid period and contains references to foreign ships that entered Egypt — most certainly

via Heracleion. According to this document larger boats had a tonnage of 60 tonnes and smaller

ones of 40 tonnes.37 Such igures, for instance, seem to correspond well to the dimensions of the

ships of Heracleion.

4. The Context of the Shipwrecks

4.1. Position of the Shipwrecks

Two distinctive accumulations of ships are situated in the center of harbour H3. What

could be the reason for this situation? Little can be ascertained for the time being but several

facts are worth mentioning. First, it seems that there is not much material on board of the ships

and except one shipwreck loaded with limestone blocks there is no general cargo. A strange fact,

keeping in mind that these trade ships were anchored in the harbour of the largest emporion of

the ancient Mediterranean. If a sudden natural disaster would have sunk them at least some of

the ships would have been partially loaded. he same shipwreck 17 that is located in the pas-

sage has a well-preserved hull but not a single artefact could be deinitely associated with it. On

the other hand, the analysis of the hull’s internal structure (see above) showed that some details

are missing and only the largest elements are still in place. Secondly, shipwreck 17 is surround-

ed by wooden posts and other posts were found between the earlier shipwrecks 44 and 45. It

35 Fabre in Wilson and Robinson, East Meets West.36 J. Rougé, La marine dans l’Antiquité (Paris, 1975).

37 Porten and Yardeni, Textbook of Aramaic Doc-uments, III.


seems possible that several posts could have been used to moor the ships that waited their turn

to be loaded or unloaded, although this is rather inconvenient for navigation, and sometimes

even dangerous. Using anchors would have been a preferable option. In fact, surrounding ships

with posts can be explained in two ways only. Either it meant that the ships were intentionally

positioned in order to block the entrance to some strategically important point or that they

were used as loating peers. he irst possibility is well documented in naval practice,38 though

the second one is also well-known.39

Some of the shipwrecks can perhaps be associated with the anchors found nearby. Several

anchors are positioned within the contours of the hull and could have been on board of the ships

when they sank. he anchor from shipwreck 43 is a special case as it seems to be in the ready

to use position at the bows of the ship (Fig. 6). he number of anchors found at our site is not

surprising as stone anchors were lost easily and each ship could have carried a dozen of them on

board. More interesting is the distribution of anchors by type. his typology may provide much

information about navigation practices in the Heracleion’s area and it is currently being studied.

4.2. Artefacts Found on Board

he ceramic objects recovered from the shipwrecks date from the late 5th to the 4th century bc.

Of special note is the discovery between shipwrecks 4 and 6 of a gold coin (H3_10717)

with the depiction of Zeus. A Greek inscription indicates that the coin was minted by Pix-

odarios, the Persian satrap of Caria (340–335 bc). Another interesting fact is that in the east

accumulation of the shipwrecks, a Greek helmet was discovered dating to the 5th–4th centuries bc

(HXX_8295, SCA 1026).40

4.3. Historical Context of the Shipwrecks

For the moment it is too early to develop hypotheses concerning the historical events that

led to the sinking of the boats at Heracleion-honis. To name but just a few of the events that

took place in the Delta region in the 5th–4th centuries bc we must mention the Inaros revolt and

the Athenian intervention in 466 bc41, the reconquest of Egypt by the Persians (343–342 bc)42

or the reign of Khabbabash during the intermediate period between 343 and 336–335 bc. All

these events were connected with the Canopic region and Heracleion-honis which occupied

a strategic area at the entrance to the Nile (and Egypt). he ships of Heracleion-honis could

have been sunk for various reasons: in a naval battle, intentionally or in accordance with a de-

fence strategy, in order to destroy enemies transports or as a punishment.43 Further excavations

will allow for a deeper understanding of both the features of the construction of the ships and

their place in Egyptian history.

38 From recent times we can refer here to the siege of Sebastopol during the Crimean War (1854–1855) when an entire leet was sunk to block access to the harbour.

39 Pomey in 6th International Symposium on Ship Construction in Antiquity, 425–437.

40 Secunda in Goddio and Fabre, Egypt’s Sunk-en Treasures, 336 nr. 307.

41 Diodorus, Bibliotheca historica XI, 71, 3–6.42 Diodorus, Bibliotheca historica XVI, 40, 6.43 he historical and geographical context of

the shipwrecks is presented by Fabre in Wilson and Robinson, East Meets West.


Figure 1. Calibrated dating C14 of 45 shipwrecks of Heracleion-honis and their dating probability curve (%)

Figure 2. Calibrated dating C14 of two accumulations of ships in the central harbour of Heracleion-honis and their dating probability curve

Figure 3. Percentage of ships as per the wood of their planking (one sample per ship) analyzed by Archeolabs and the IFAO

Th

e Shipw

rec

ks o

f Her

acleio

n-T

ho

nis: A

n O

ver

view

117

Figure 4. Upper part of the preserved planking of shipwreck 21 (Acacia sp., conventional 14C dating 540 +/- 50 bc, calibrated C14 dating 787 cal bc–482 cal bc) and joint reconstruction (right). Patrice Sandrin © Franck Goddio, Hilti Foundation


Figure 5. Limestone blocks (ballast ?) and Shipwreck 3 (Acacia sp., conventional C14 dating 300 +/- 50 bc, calibrated C14 dating 398 cal bc–184 cal bc).

Heracleion, Port H3. Christoph Gerigk © Franck Goddio, Hilti Foundation.

Figure 6. Anchor found on shipwreck 43 with its position on excavation plan. Limestone and wood (Pinus sp., C14 calibrated date: 405 cal bc–208 cal bc), l. 75 cm, w. 50 cm, th. 18 cm.

Heracleion, Port H3. Christoph Gerigk / Patrice Sandrin © Franck Goddio, Hilti Foundation.